Anthropic Donates $20 Million to Public First Action for AI Governance

Anthropic has contributed $20 million to Public First Action, a bipartisan 501(c)(4) organization. This funding is intended to support public education on artificial intelligence, promote the implementation of AI safeguards, and ensure the United States maintains its lead in the global AI race.

AI Risk Landscape and the Need for Policy

Anthropic identifies a critical need for flexible regulation to manage the risks associated with rapidly advancing AI capabilities. The company notes that AI is already being used to automate cyberattacks and could potentially assist in the production of dangerous weapons. Additionally, powerful AI systems may take harmful actions that are contrary to the intentions of their users.

The pace of AI development is described as "dizzying," moving from simple chatbots in 2023 to complex agents capable of completing intricate tasks. Anthropic highlights that this rapid progress impacts various professions, including software engineering, where the company has had to redesign technical hiring tests multiple times as AI models defeated previous versions.

Public First Action: Goals and Bipartisan Approach

Public First Action is a bipartisan organization led by both Republican and Democratic strategists. It aims to fill a gap in organized efforts to mobilize politicians and the public regarding AI governance. This initiative comes as polling indicates that 69% of Americans believe the government is not doing enough to regulate AI.

Public First Action focuses on four primary policy priorities:

• AI Model Transparency: Insisting on safeguards that provide the public with more visibility into how frontier AI companies manage risks.
• Federal Governance Framework: Supporting a robust federal AI governance framework while opposing the preemption of state laws unless Congress enacts stronger safeguards.
• Export Controls: Supporting smart export controls on AI chips to prevent authoritarian adversaries from accessing critical technology.
• Targeted Regulation: Pursuing regulation focused on immediate high-risk areas, specifically AI-enabled biological weapons and cyberattacks.

Scope and Intent of Governance

Anthropic states that these policy goals are not intended to benefit the company as a developer, as effective governance would result in increased scrutiny of AI labs. Furthermore, the company specifies that transparency regulations should apply specifically to companies developing the most powerful and dangerous AI models, rather than smaller or less well-resourced developers.

Anthropic frames this contribution as part of its responsibility to ensure that AI technology serves the public good and manages risks proportionately to the technology's transformative potential.